CentralTendency-UserGuide.md

PhUSE CS Central Tendency Standard Analyses

User & Contributor Guide

Version 0.1

PhUSE CS Central Tendency analyses and displays are described in detail in the White Paper on Measures of Central Tendency (WPCT), which PhUSE have published in the CS Deliverables Catalog

Scope

Analyses and their SAS and R implementations are based on ADaM-compliant data sets.

Out of scope:

• ADaM conformance checks
• Analyses of tabulated data, SDTM

Participating & Contributing: Getting started

We maintain to-do lists in project directories:

• Working Group 5 to-do list
• WPCT to-do list

Review the Script Basics and Programming Conventions, below.

SAS and R versions

• Standard SAS scripts use functionality in SAS 9.4 M02 or later, utilizing ODS and GTL. See our visual guide to the WPCT PhUSEboxplot GTL template. (For users limited to SAS 9.2, we provide a SAS 9.2 boxplot approach in the script WPCT-F.07.01-sas92-QCshewhart.sas)

• Standard R scripts use functionality in the current R release and corresponding R packages.

WPCT Package Scripts & Conventions

Each Central Tendency data display will have a same-name script in SAS and R in the WPCT folder. For example, SAS and R scripts for Fig. 7.1, Fig. 7.2, etc.

Script basics

• We refer to these as "Standard Scripts" because they produce one of the standard displays enumerated in the white paper.

• Standard scripts, by default, produce an example output based on PhUSE CS test data.

  • CDISC publish these data on their website
  • We have modified the CDISC data to enhance script testing, and publish these modified data in github

• Standard SAS scripts require our library of macros, which we publish here in github

• Standard R scripts require some common functions the R functions will be created in 2017.

• Scripts require that users set some parameters based on their particular computing environment and data

  • Scripts are organized to clearly isolate user settings and any pre-processing instructions (e.g., to subset data before performing analyses)
  • By default, scripts use the CDISC ADaM pilot data mentioned above, and include custom pre-processing code to abbreviate treatment labels and subset data to limit the number of outputs.

Project conventions

• The whitepapers README file describes our directory structure.

• Programming Guidelines for this project are in our PhUSE Wiki

• We outline the Qualification Process for this project in our PhUSE Wiki, as well

• We publish Test (Qualification) plans, scripts and results in GitHub

• We publish an Index of Standard Scripts in our GitHub Wiki

• We also publish separately an Index of SAS Macros in our GitHub Wiki. The SAS Standard Scripts requires these macros.